Do I need to RMA this card?

[fpoverflow]

Hi all,

I recently purchased a sapphire x1950xt card. This one, to be specific:

http://www.newegg.com/Product/Product.aspx?Item=N82E16814102067

Whenever I play Supreme Commander, the game eventually crashes. I have ATItool running now, but before I was using that the game would crash after 5-10 minutes without fail. With ATItool and the fan settings set very aggressive (100% 69C and up), the game will run for 30-60 minutes. Of course, my PC sounds like an aircraft engine. The log file shows temps consistently between 70-75C when playing the game, with ATItool blasting the fans.

My PS is a 620w corsair, so I don't think that is the problem in this case.

Is this normal behavior? Should I RMA this card in your opinion?

Thanks for your input.

 

[mruffin75]

Do you have any extra cooling in your case? No matter how good the PSU is, if your case is sealed shut with no ventilation, the video card will overheat..

 

[manimal]

Open the side of the case and play and see if it still crashes. Run ati tool and scan for artifacts, check your PSU amperage and make sure its enough to run your card at load.

 

[fpoverflow]

I have 2x120mm fans going. One for intake and exhaust. Two of these: http://www.newegg.com/product/product.asp?item=N82E16811998127

This is the power supply. Should be powerful enough: http://www.newegg.com/Product/Product.aspx?Item=N82E16817139002

+3.3@24A,+5V@30A,+12V1@18A,+12V2@18A,+12V3@18A,- 12V@0.8A,+5VSB@3A

I tried to find the max core with atitool, but it crashed the system.

I'll give the scan artifacts a try, and also opening up the case. The GPU env temp reading is 49C at idle. Processor is at 35C.

 

[fpoverflow]

Ok, I've been running scan artifacts for 8m now. The temp on the GPU is at 80C, the fan is 100%. No errors yet though. Probably if I let it keep going it would crash in about 30m. The GPU env temp is pretty stable at 58C. It seems to have topped off at 80C for now. Amps are pretty stable at 30A according to atitool at well.

update: I opened up the case. Oddly enough, the gpu env is stable at 60C, so it didn't drop. And gpu temp itself stayed at 79C, so opening the case didn't make much of a difference at all. 19m of scan artifacts and no errors till though.

 

[fpoverflow]

Ultimately, I ran scan artifacts for 1.5 hours without problem. Is that a decisive test?

I'm curious why SC crashed so quickly (repeatedly) when the fan settings were at stock, but not as quickly when the fans were cranked up. It really points to an overheating issue, but the scan artifacts indicates otherwise.

 

[Capt Caveman]

What drivers are you using?

 

[fpoverflow]

This is the info from Catalyst

Driver Packaging Version 8.36-070314a3-044981C-ATI
Catalyst® Version 07.4
Provider ATI Technologies Inc.
2D Driver Version 6.14.10.6683
2D Driver File Path System/CurrentControlSet/Control/Video/{33632A70-FE36-4C92-BB4A-E5BF67E6DC5E}/0000
Direct3D Version 6.14.10.0489
OpenGL Version 6.14.10.6458
Catalyst® Control Center Version 2007.0322.2238.38535
AV Stream (T200) Driver Version 6.14.10.1078

 

[apoppin]

you need to compare notes with this guy:

http://forums.anandtech.com/messageview...atid=31&threadid=2048415&enterthread=y

he had 2 xts with problems similar to yours

--same games i think, and even the same PS !!

 

[fpoverflow]

Interesting. I think I'll try loosening my mem timings a bit, and perhaps reseat the heatsink and put some AS5 on it, if the cleanup isn't too involved. I'm assuming doing so wouldn't prevent me from making an RMA if necessary.

 

[Captante]

Originally posted by: fpoverflow
Interesting. I think I'll try loosening my mem timings a bit, and perhaps reseat the heatsink and put some AS5 on it, if the cleanup isn't too involved. I'm assuming doing so wouldn't prevent me from making an RMA if necessary.

You might be able to get away with RMA'ing the card anyway, but removing the factory HS/fan from a Sapphire voids your warranty.

 

[tigersty1e]

Originally posted by: fpoverflow


update: I opened up the case. Oddly enough, the gpu env is stable at 60C, so it didn't drop. And gpu temp itself stayed at 79C, so opening the case didn't make much of a difference at all. 19m of scan artifacts and no errors till though.

You have to blow a fan on the open case otherwise it'll be the same.

 

[HybridSquirrel]

hey buddy you got a solution yet?

now my card gives me dx9 errors in games, than crashes on me. =[

 

[fpoverflow]

Not anything definitive yet. My problem seems a little different than yours. What will happen to me is that SC will crash itself and generate an error report, and then I'll end up back in windows. Only occasionally does the system BSOD. I noticed last night that I had my PCIe overclock to 105, so I put that back down to 100 to see if that will make any difference. I'm sure you're already doing this, but crank the fan to 100% at 60C using ATItool, that helped me a fair bit.

 

[HybridSquirrel]

yeah mines more at 65c its 100%

alot of times the game generates error reports for me aswell but yeah i had my pcie to 100 so i set that to auto which when im not overclocked should be 100, but that worked pretty well for me. also it doesnt hurt to rma and try out a new card. alot of times if you talk to newegg llive chat they will send you a label in the mail